phenotype of columnar-lined esophagus in rats with esophagogastroduodenal anastomosis: similarity to...

Phenotype of columnar-lined esophagusin rats with esophagogastroduodenalanastomosis: similarity to humanBarrett’s esophagus

Yinghao Su1,*, Xiaoxin Chen2,*, Michael Klein1, Ming Fang1, Su Wang2, Chung S Yang2

and Raj K Goyal1

1Center for Swallowing and Motility Disorders, VA Boston Healthcare System and Harvard Medical School,Boston, MA, USA and 2Susan Lehman Cullman Laboratory for Cancer Research, Department of ChemicalBiology, Ernest Mario School of Pharmacy, Rutgers, The State University of New Jersey, Piscataway, NJ, USA

In rats, esophagogastroduodenal anastomosis (EGDA) without concomitant chemical carcinogen treatment canlead to columnar-lined esophagus (CLE) including metaplasia, dysplasia, and esophageal adenocarcinoma(EAC). This study describes the morphology and phenotypic features of CLE and EAC in the rat model andcompares them with the corresponding lesions in human Barrett’s esophagus (BE). Swiss roll preparations ofesophagi of EGDA rats and biopsies from human BE containing specialized intestinal metaplasia (SIM) and EACwere examined. The esophagi of EGDA rats showed esophagitis, CLE, islands of multilayered epithelium (MLE),dysplasia and EAC. The CLE had features of specialized intestinal metaplasia. MLE frequently occurred at theneo-squamocolumnar junction and occasionally in the mid-esophagus in isolated foci. Scattered mucinouscells in esophageal squamous epithelium were also found. The CLE and MLE in EGDA rats resembled thelesions described in human BE in morphology, mucin features and expression of differentiation markers (CK7,CK20, Das-1, villin, and pS2/TFF1). Invasive EAC in EGDA rat is of well-differentiated mucinous type, which is incontrast to the variably differentiated glandular type of adenocarcinoma in human BE. p53, c-myc, andcyclooxygenase 2 are expressed in both the rat and human SIM and EAC. These studies indicate that, notwithstanding small differences, SIM and EAC induced in EGDA rats are similar to the corresponding lesions inhuman BE. EGDA rats may serve as a useful model to study the pathogenesis, molecular biology, andchemopreventive interventions of human BE and EAC.Laboratory Investigation (2004) 84, 753–765, advance online publication, 19 April 2004; doi:10.1038/labinvest.3700079

Keywords: intestinal metaplasia; esophageal adenocarcinoma; phenotype of columnar-lined esophagus

Esophageal adenocarcinoma (EAC) is one of thefastest rising cancers in the Western countries.1 Thereasons for this alarming increase are not fullyunderstood. Barrett’s esophagus (BE) is the mainlesion underlying the development of EAC. BE isdefined as the replacement of normal squamousepithelium by columnar epithelium, which includesspecialized intestinal metaplasia (SIM) as its majorconstituent.2 SIM is characterized by features of type

III intestinal metaplasia and is the lesion that isassociated with the development of EAC. Theestimated risk of EAC in BE patients with SIM isaround 0.5% per year which is 30- to 125-foldhigher than the general population.3 BE is thought tobe a complication of gastroesophageal reflux dis-ease, and the development of BE may explain thehigh association of gastroesophageal reflux diseaseand EAC.1 Despite the clinical importance of BE andEAC, the etiology, cellular origin and pathogenesisof BE and SIM are not well understood.4 The roles ofgastric acid, bile and other gastrointestinal contentsin the pathogenesis of BE and EAC also remainunclear. Availability of an animal model may helpaddress some of these issues.

Squamous cell carcinoma, adenocarcinoma, andmixed adenosquamous carcinoma are three types ofcancer that can be induced in experimental animals.

Received 06 November 2003; revised 21 January 2004; accepted25 January 2004; published online 19 April 2004

Correspondence: RK Goyal, MD, VA Medical Center, 1400 VFWParkway, West Roxbury, MA 02132, USA.E-mail: [email protected]

This work is supported by NIH Grants DK031092 (RK Goyal),

DK63650 (X Chen) and CA75683 (CS Yang).*Contributed equally to this publication.

Laboratory Investigation (2004) 84, 753–765& 2004 USCAP, Inc All rights reserved 0023-6837/04 $25.00

www.laboratoryinvestigation.org

Carcinogens such as 2,6-dimethylnitrosomorphineor methyl-n-amylnitrosamine induce esophagealsquamous cell carcinoma in experimental animals.On the other hand, gastroduodenal contents alsoappear to be carcinogenic. Surgical procedures thatpromote gastroduodenal reflux lead to columnarmetaplasia and esophageal adenocarcinomas.5–7

When chemical carcinogens are combined withsurgically induced gastroesophageal reflux of gastro-intestinal contents, mixed adenosquamous carcino-mas are generally produced.8,9

Goldstein et al10 reported that rats with esopha-goduodenal anastomosis develop EAC withoutadenosquamous or squamous cell carcinoma, andthe frequency of carcinoma development increasesby correction of postoperative anemia with parent-eral iron therapy. Similarly, rats surgically treatedwith esophagogastroduodenal anastomosis (EGDA)to induce esophageal reflux of gastric acid and biledeveloped columnar-lined esophagus (CLE) andEAC in over half of the animals by 40 weeks aftersurgery.11 The morphological and phenotypic fea-tures of the experimental CLE have not been welldescribed. Furthermore, the resemblance of thismodel to human BE and EAC remains unclear.12

Pathological characterization is crucial to theidentification of animal models of human disease.13

In this study, we describe the morphologicalfeatures, mucin histochemistry, expression patternof differentiation markers (CK7, CK20, Das-1, villin,and pS2/TFF1) and cancer-related gene products(p53, c-myc, and cyclooxygenase 2) in CLE and EACin the EGDA rat model. We then compare the EGDArat model with findings in human tissues. Ourstudies show that, in spite of a few differences, thereare remarkable similarities of SIM and EAC betweenthe EGDA rats and the human diseases. The EGDArats may serve as a useful experimental model tostudy the pathogenesis and cellular origin of humanBE. This model may also help in identifying theintermediate molecular markers and developingtreatment strategies for human BE and EAC.

Materials and methods

Animal Models and Tissue Preparation

The EGDA rats were prepared as previously de-scribed.11 Briefly, an anastomosis was made betweenthe gastroesophageal junction and the duodenum onthe antimesenteric border with accurate mucosal tomucosal opposition in 8-week-old male Sprague–Dawley rats. The animals were treated with iron(4 mg/kg/week, i.p.) without concomitant carcino-gens. They were kept for 40 weeks after surgery andkilled by CO2 asphyxiation. The esophagus wasremoved, fixed in 10% buffered formalin, Swiss-rolled, and processed in paraffin. Tissue samplesfrom 20 rats showing both esophagitis and CLE wereused for this study. Of them, 14 had visibleadenocarcinoma. In total, 10 nonoperated normal

rats were used as controls. The Animal Care andFacilities Committee, Rutgers University (protocol#94-017), approved the protocol for the animalstudies. For comparative studies in humans, coded(with all identifiers removed) archival paraffin-embedded biopsies having SIM and EAC wereobtained from the Department of Pathology at theVeterans Affairs Boston Healthcare System. Thestudy protocol was approved by the InstitutionalReview Board of the VA Boston Healthcare System(protocol #1546). Nine tissues had only BE and fivehad only EAC. Six tissues had both BE and EAC.

Histopathology

Tissue sections were stained with hematoxylin andeosin (H&E) for histopathological analysis. Inflam-mation was graded as mild, moderate, and severebased on the degree of infiltration by the inflam-matory cells, particularly neutrophils, eosinophils,macrophages, and plasma cells. The columnarepithelium was graded as normal, reactive orregenerative epithelium, low-grade dysplasia(LGD), and high-grade dysplasia (HGD).14 Severesquamous hyperplasia was diagnosed when surfaceprojections of subepithelial papillae caused irregu-lar surface ridges. Inverted, papilloma-like lesionsappeared as projections and infolding of the squa-mous epithelium, without neoplastic changes, intothe submucosa and the muscle layer of the esopha-gus. If small clusters of dysplastic squamousepithelial cells penetrated through the basementmembrane into lamina propria and adjacent tissues,the diagnosis of invasive squamous cell carcinomawas made.15 However, no case of invasive squamouscell carcinoma was found.

SIM was diagnosed if intestinal columnar epithe-lium fulfilled features of specialized intestinalmetaplasia and contained intestinal-type gobletcells above the anastomotic site marked by the blueprolene suture. Multilayered epithelium (MLE) wasdiagnosed when the esophageal epithelium showedlayers of both squamous and columnar epithelium.16

Dysplasia of the columnar mucosa was classified asLGD and HGD, according to the criteria suggested byHaggitt.17 In LGD, the crypt architecture trends to bepreserved and distortion is minimal; the nuclei maybe stratified, but the stratification does not reach theapical surface of the glands; nuclei are enlarged,crowed, and hyperchromatic; mitotic figures may bepresent in the upper portion of the crypt; goblet andcolumnar cell mucus is usually diminished orabsent, but goblet cells in which the mucous dropletdoes not communicate with the luminal surface maybe observed. The abnormalities extend to themucosal surface. In HGD, distortion of crypt archi-tecture usually is present and may be marked. It iscomposed of branching and lateral budding ofcrypts, a villiform configuration of the mucosalsurface or intraglandular bridging of epithelium to

Animal model of Barrett’s esophagusY Su et al

754

Laboratory Investigation (2004) 84, 753–765

form a cribriform pattern of ‘back-to-back’ glands.Nuclear abnormalities are present as in LGD, andstratification reaches the crypt luminal surface.There may be a loss of nuclear polarity, and nucleioften vary markedly in size, shape, and stainingcharacteristics. Goblet and columnar cell mucus isusually absent. The abnormalities extend to themucosal surface. EAC was diagnosed when dysplas-tic columnar epithelial cells invaded through theepithelial basement membrane and into deepertissue.

Histochemistry

Alcian blue/periodic acid Schiff (AB/PAS) and highiron diamine/Alcian blue (HID/AB) were used formucin staining. Intestinal metaplasia was classifiedas complete (type I) or incomplete (type II or typeIII). All SIM had goblet cells that containedsialomucins and/or sulfomucins. However, in typeI intestinal metaplasia, columnar cells had absorp-tive features but contained no acid mucin; in type II,columnar cells contained sialomucins (stainedblue); and in type III intestinal metaplasia, thecolumnar cells contained sulfomucin (stained pur-ple).18 Silver staining (Reticulum Gomori’s SilverKit, Newcomer Supply, Middleton, WI, USA) wasused to display the basement membrane of theesophageal epithelium.

Immunohistochemistry

The paraffin-embedded sections were deparaffi-nized, rehydrated, and pretreated by 0.1%. trypsinat 371C for 30 min or by heating slides for 5–10 minin 10 mM citrate buffer. The staining was performedwith the ARKt Peroxidase kit (Dako, Capenteria,CA, USA) according to the manufacturer’s instruc-tions. The sources and working concentrations ofprimary antibodies are as follows: CK7 (4.1 mg/ml,Clone OV-TL 12/30, DAKO), CK20 (1.2 mg/ml, CloneKs 20.8, DAKO), Das-1 (1:10, a gift kindly providedby Dr Kiron M Das, Department of Medicine,Molecular Genetics and Microbiology, Universityof Medicine and Dentistry of New Jersey), villin(1.3 mg/ml, Clone CWWB1, NeoMarkers, Fremont,CA, USA), pS2/TFF1 (5 mg/ml, Innogenex, SanRamon, CA, USA), p53 (0.4 mg/ml, Clone PAB 240,Chemicon, Temecula, CA, USA), c-myc (8 mg/ml,Clone C-8, NeoMarkers), and COX-2 (2.5 mg/ml,Transduction Lab, Los Angeles, CA, USA). Fornegative controls, normal serum or phosphate-buffered saline (PBS) replaced primary antibodies.Both positive and negative control slides wereprocessed in parallel.

The staining intensity was subjectively gradedinto four degrees, from �, þ , þ þ , to þ þ þaccording to the following criteria. �, no positivestaining; þ , less than 2/3 cells with slight stainingor less than 1/3 cells with slight to moderate

staining; þ þ , more than 2/3 cells with slightstaining or 1/3 to 2/3 cells with slight to moderatestaining; þ þ þ , more than 2/3 cells with moderateor strong staining. Statistical analysis was con-ducted with Fisher’s exact test for frequency dataand Ridit analysis for ordinal data.19

Results

Normal Esophageal Mucosa

Normal esophageal epithelium in the rat is stratifiedsquamous, consisting of a single layer of basal cellswith round nuclei, four to six layers of prickle cells,and a thin stratum corneum on the surface. Distallythe squamous epithelium is continuous with thesquamous epithelium of the forestomach. The ratesophagus has no submucosal glands. This is incontrast to the normal human esophageal epithe-lium that has stratified squamous epithelium with-out keratinization and also has submucosal glands.Moreover, in humans, distally the squamous epithe-lium becomes continuous with the columnar gastricepithelium forming the squamocolumnar Z line.1

Columnar Metaplasia in the Esophagus

Columnar mucosa is not present in the normalesophagus. Replacement of the normal squamousepithelium with metaplastic columnar epitheliumfrequently occurred in the EGDA rats. The columnarmetaplasia appeared either as a single type CLE or asMLE (columnar and squamous) (Figure 1). Thesepatterns are similar to those found in the human BE.

CLE and Specialized Intestinal MetaplasiaOn H&E staining, the CLE in EGDA rats consist of asingle layer of surface columnar epithelium andtubular mucosal glands. Columnar cells have fea-tures of both mucous secretory and absorptive cellswith interspersed intestinal-type goblet cells. CLE ispresent in the distal esophagus near the anastomoticsite and is bound by the squamous epithelium of theesophagus and fore-stomach, and the columnarepithelium of the anastomosed duodenum.

In order to define whether the columnar mucosain the lower esophagus represented the anasto-mosed duodenal mucosa or metaplastic columnarepithelium of esophageal origin, we examined theneo-squamocolumnar junction and the anastomoticjunctions for the continuity of the surface epitheliaand the basement membrane. At the anastomoticsite, the duodenal and the esophageal or the gastricsquamous epithelia as well as their basementmembranes were discontinuous. On the other hand,the mucosal junction of the esophageal squamousepithelium and the CLE were continuous withoutinterruption of basement membrane (Figure 2).Moreover, the morphological features of the duo-denal mucosa were quite distinct from those of the


755


CLE. These observations are consistent with theview that the columnar mucosa lining the esophaguswas metaplastic epithelium of esophageal origin. Insome cases, CLE was covered by squamous epithe-lium as has been described in humans, and may berelated to partial healing of BE. All of the 20 cases ofCLE had intestinal metaplasia, including 17 cases oftype III and three cases of type I. There were 14cases of dysplasia, including 12 LGD and two HGD(Table 1).

On mucin staining (PAS/AB and HID/AB), CLE inthe EGDA rats shows that the goblet cells containedboth sialomucin and sulfomucin, whereas thecolumnar cells stain for neutral mucin, sialomucinor sulfomucin. The staining intensity of sulfomucinis similar to that of sialomucin in 15 of 20 cases. Themucin stains clearly distinguished CLE from the ratduodenal mucosa, which showed sulfomucinin the goblet cells but not the columnar cells.20

However, the mucin staining features of CLE in therat (Figure 3a) and human BE (Figure 3b) weresimilar, resembling the type III intestinal metaplasia

in most cases. In order to characterize further thephenotype of SIM in the EGDA rats, we examinedfor the presence of several other markers andcompared the results with similar studies in humanBE (Table 2). The CK7/CK20 pattern of superficialCK20 staining in surface glands and diffuse CK7staining in surface and deep glands has beenattributed to be specific to BE.21 This pattern wasobserved in three of 20 rat CLE (Figure 3c, e) and 11of 15 human BE (Figure 3d, f) (Fisher’s: Po0.05).Other staining patterns, such as diffuse CK7/diffuseCK20 (one rat and one human), superficial CK7/superficial CK20 (one rat and one human), anddiffuse CK7/negative CK20 (one rat and two hu-mans) were also observed. Four cases of rat SIM haddiffuse expression of CK20 and some scattered CK7-positive cells. The remaining 10 cases of rat SIM didnot express either CK7 or CK20.

Das-1, a polyclonal antibody, which recognizesnormal human colon epithelium and BE in thegastrointestinal tract,22 positively stained nine of 20rat CLE and 14 of 15 human BE (Fisher’s: Po0.05).

Figure 1 CLE and MLE in an EGDA rat. The figure is montage of a swiss-roll preparation of an EGDA rat esophagus that has been stainedwith AB/PAS. CLE is demarcated by dashed lines and appears as an area of extensive intestinal metaplasia at the distal end of theesophagus. Note that CLE is continuous with squamous epithelium of the forestomach (r). MLE is demarcated by dotted area in the mid-esophagus and appears as island of metaplastic epithelium. MLE containing a mixture of squamous and columnar epithelium can also beidentified at the neo-squamocolumnar junction (a). Changes in the squamous epithelium including hyperplasia (D), papillary projections(m), inverted papilloma-like lesions (.), erosions (b), and sloughing (g) are frequently observed in the EGDA rats. Scale bar: 200 mm.


756


The cytoplasmic immunostaining of Das-1 wasweaker in the rat CLE (Figure 3g) than in humanBE (Figure 3h). Villin, a cell differentiation markerfor intestinal brush-border formation,23,24 was ob-served frequently in rat SIM than in human SIMwithout significant difference (Fischer’s: P40.05).

The cytoplasmic immunostaining of villin wasusually stronger in the rat (Figure 3i) than in thehuman (Figure 3j) tissues. The cytoplasmic stainingof pS2/TFF1, a marker of ulcer-associated celllineage,25 was found in 17 of 20 rat CLE (Figure3k) and 13 of 15 human BE (Figure 3l) with similarfrequency (Fisher’s: P40.05) and immunoreactivity.

Multilayered epitheliumIn three cases of EGDA rats, scattered mucinouscells appeared admixed with hyperplastic squa-mous epithelium (Figure 4a). In three other cases,MLE was identified as islands of mucinous cellsinterspersed in squamous epithelium (Figure 4b).The columnar and goblet cells may appear on thetop, in the middle, or at the bottom of the squamouscell component of MLE. One case had mucinouscells located at the superficial squamous epitheliumfar removed from the anastomosis site (Figure 1).Two cases had mucinous cells located in deep nestsof hyperplastic squamous epithelium in the neo-squamocolumnar junction. Multiple foci of MLEwere found in one case and presented as threeseparate foci distributed in distinct nests of hyper-plastic squamous epithelium. In our biopsies ofhuman tissue, MLE was not found. This isnot surprising because this material representeddirected biopsies of the CLE. In humans, MLE isfound only in random biopsies taken at thesquamocolumnar junction. The morphologic featureof MLE in the rat model closely resemble thosedescribed in the human patients.16,26

Figure 2 Showing metaplastic nature of the columnar mucosa in EGDA rats. (a) H&E stain of the neo-squamocolumnar junction at thedistal end of the esophagus in an EGDA rat. Note that the metaplastic columnar epithelium (-) is interposed between areas of squamousepithelium on either side. (b) Gomori silver-Alcian blue-hematoxylin triple staining showing that the basement membrane (short blackarrow) between the squamous and columnar epithelia is not interrupted, thus arguing against the mucosal junction being a surgicalapproximation of duodenal and esophageal mucosa. (c) Represents villin immunostaining of the neo-squamocolumnar junction. Thecolumnar epithelium shows villin immunoreactivity (dark brown), and squamous epithelium is negative for villin immunostaining. Notethe direct continuity of the metaplastic columnar epithelium with the squamous epithelium. Scale bars: 100mm.

Table 1 Histopathology of esophageal mucosa in the rat modeland human BE

Rat model Human BENo. of cases

(%)No. of cases

(%)

Total no. examined 20 15Squamous epitheliuma

Regenerative 4 (20) 6 (50)c

Low-grade dysplasia 12 (60) 6 (50)High-grade dysplasia 4 (20) 0Surface ridges 2 (10) 0Inverted papilloma-like lesion 7 (35) 0

Inflammatory responseb

None 0 0Mild 4 (20) 6 (40)Moderate 12 (60) 9 (60)Severe 4 (20) 0

Intestinal metaplasiaType I 3 (15) 3 (20)Type II 0 0Type III incomplete 17 (85) 12 (80)

DysplasiaNone 6 (30) 3 (20)LGD 12 (60) 6 (40)HGD 2 (10) 6 (40)

aRidit: m¼ 2.19 (Po0.01); bRidit: m¼ 1.8 (P40.05); cSquamous

epithelium was absent in three human tissues.


757



758


We also examined the phenotypic characters ofthe columnar cells of MLE in the rat model.Mucinous cells in the rat model contained bothsulfomucin and sialomucin, but little to no neutralmucin (Figure 4c, d). The foci of MLE were readilydetected with AB/PAS staining. The mucinous cellsof rat MLE in the deep nests and surface ofhyperplastic squamous epithelium showed strongimmunoreactivity of CK7, moderate to strong stain-ing of villin, weak staining of Das-1 and pS2/TFF1,and negative staining of CK20 (Figure 4e–h). Thesebiochemical markers in the rat are similar to thosedescribed in humans.16

Esophageal Adenocarcinoma

All of the 14 cases of EAC in EGDA rats werelocalized to the esophagus just above the esophago-gastroduodenal anastomosis. Two cancers wereearly cancers limited to the submucosa. In total, 12cases were advanced EAC showing invasion ofcancer cells into muscle, adventitia, and adjacenttissues (Figure 5a–c). These cancers clearly arosefrom the SIM in the esophagus rather than from the

duodenal mucosa (Figure 5d). EACs in the rat modelwere highly differentiated mucinous carcinomas. Incontrast, 11 cases of human EAC showed variousdegrees of differentiation (well—46%, moderately—27%, and poorly—27%) with glandular structuresin 10 cases and mucinous morphology in only onecase.

Similar to adjacent CLE, rat EAC had largequantities of sulfomucin and sialomucin, and lesseramounts of neutral mucin (Figure 5e). CK7 (Figure5f), CK20, and Das-1 were less frequently expressedin rat EAC than in human EAC (Table 2). Rat EACalso expressed villin (Figure 5g) and pS2/TFF1.

Accumulation of the tumor suppressor p53 pro-tein was found in 29% (4/14) of EAC and 75% (15/20) of CLE in EGDA rats (Fisher’s: Po0.05). The p53protein was localized in the nuclei and cytoplasm,mainly in the deep glands of the neo-squamocolum-nar junction (Figure 6a). Cytoplasmic staining ofoncogene c-myc was found in the CLE (5/20, 25%),but not EAC, of EGDA rats (Figure 6b). Theepithelial COX-2 immunoreactivity was seen morefrequently in rat CLE than in human BE (Figure 6c)(Table 2). Hyperplastic squamous epithelium, fibro-blasts, and infiltrated inflammatory cells alsoshowed weak to strong COX-2 immunoreactivity inboth EGDA rats and human patients.

Changes in Esophageal Squamous Epithelium in theRat Model

Esophageal squamous epithelium of EGDA ratsexhibited various degrees of mucosal erosion,ulceration, and infiltration of acute inflammatorycells, both eosinophils and neutrophils. Basal cellhyperplasia was so severe that it often occupied theentire thickness of the mucosa. Projections offibrovascular papillae with a thin core formedmucosal ridges. A prominent feature was invertedsquamous papilloma-like lesions in which rete pegsextended down into the submucosa to the muscle(Figure 1). Inverted papilloma-like lesions have notpreviously been described in the esophagi of hu-mans or experimental animals, but have been welldescribed in the nasal mucosa and the urinarybladder.27 Single or multiple inverted papilloma-like lesions were found in seven out of 20 EGDArats. Unlike squamous cell carcinoma, inverted

Figure 3 Comparison of mucin staining and immunohistochemical features of rat CLE and human BE. (a, b) HID/AB staining show thatgoblet cells of CLE in EGDA rats (a) contain both sialomucin (blue) and sulfomucin (dark brown or black), whereas columnar cells stainfor sulfomucin. This feature is similar to that seen in human BE (b). (c–f) show CK7/CK20 immunostaining in rat CLE and human disease.In the rat CLE, CK7 immunoreactivity is present in both the superficial crypts and the deep gland epithelium (c); CK20 immunoreactivityis present only in the superficial epithelium and crypts (e). This CK7/CK20 immunostaining pattern is similar to the BE-specific CK7/CK20 pattern in human BE (d, f). (g, h) show Das-1 immunostaining which was seen in both the animal model and in human tissues.However, cytoplasmic staining of Das-1 is weaker in EGDA rats (g) than in human BE (h). As shown in (i, j) cytoplasmic staining of villinis also observed in the metaplastic epithelium in EGDA rats as well as in human BE tissues. However, villin staining is stronger in EGDArats (i) than in human BE (j). A similar expression pattern of pS2/TFF1 was found in rat CLE (k) and human BE (l). In both cases,superficial and upper glandular epithelium shows strong immunostaining of pS2/TFF1. Scale bars: (a, b, g and h) 100mm; (c–f, i, j)150mm; (k, l) 200mm.

Table 2 Differentiation and cancer-related markers in interstitialmetaplasia and adenocarcinoma in the rat model and human BE

SIM No. of cases (%) EAC No. of cases (%)

Ratmodel

HumanBE

Ratmodel

HumanBE

Total no. of tissues 20 15 14 11Differentiation markersCK7 10 (50) 15 (100)a 4 (29) 10 (91)a

CK20 9 (45) 13 (87)a 4 (29) 9 (82)a

CK BE pattern 3 (15) 11 (73)a NA NADAS-1 9 (45) 14 (93)a 1 (7) 8 (73)a

Villin 13 (65) 5 (33) 9 (64) 3 (27)PS2/TFF1 17 (85) 13 (87) 13 (93) 6 (55)

Cancer-related markersP53 15 (75) 10 (67) 4 (29) 3 (27)c-myc 5 (25) 2 (13) 0 3 (27)COX-2 16 (80) 3 (20)a 7 (50) 2 (18)

aFisher’s exact test, significantly different from the rat model

(Po0.05).NA, not applicable.


759



760


papillary extensions neither have irregular edgesnor dysplastic cells spreading into the adjacentstructures. No invasive squamous cell carcinomawas found. The squamous mucosa from humanpatients with BE showed mild-to-moderate basalcell hyperplasia but intramural papillary extensionswere not observed.

Discussion

This study shows that rats with esophagogastroduo-denal anastomosis (EGDA) develop: (1) SIM that hasphenotypic features similar to those found in humanBarrett’s esophagus; (2) MLE that has recently beendescribed in human disease; and (3) EAC. Thisstudy suggests that EGDA rats may serve as asuitable model for investigations of pathophysiol-ogy, molecular biology of progression to cancer, andprevention and treatment trials of human Barrett’sesophagus.

The esophageal mucosal changes in EGDA rats arelikely to be secondary to the reflux of gastroduode-nal contents into the esophagus. Although, we didnot perform quantitative measurement of gastricacid duodenal contents or bile in the esophagusof these animals, presence of free bile reflux inthe EGDA rats was evident by bile staining of theesophageal mucosa. Bile staining of esophagealmucosa was not seen in normal control. In a relatedanimal model, increased bile acids have beendescribed in the esophageal aspirate of rats afteresophagojejunostomy.6,28 Other investigators havedescribed that duodenal or gastroduodenal contents,particularly deoxycholic acid may lead to esopha-geal adenocarcinoma7 by activating NF-Kappa-Band inducing IL-8 expression.29,30 Role of gastricacid in pathogenesis of esophageal adenocarcinomais unclear.31,32

The morphology and mucin histochemistry of thecolumnar metaplasia in the rat model are identicalto those that characterize the specialized intestinalmetaplasia in human Barrett’s esophagus, includingmucous-containing columnar cells that have bothsecretory and absorptive features and intestinal typeof goblet cells that contain sulfomucins. Thesefeatures are also similar to what has been describedas type III intestinal metaplasia. The SIM in therat model also expresses phenotypic markers thatare seen in SIM in human Barrett’s esophagus,including the expression of differentiation markers

(Figure 2). The existence of BE-specific CK pattern,colon-specific antigen (Das-1), villin and a marker ofulcer associated cell lineage (pS2/TFF1)33 in themetaplastic epithelium in the rat is similar to thatseen in humans suggesting that SIM in rats andhumans may have similar cellular origin. However,differentiation markers such as CK7/CK20 and Das-1were less often expressed in the rat than in thehuman tissues. The reason for this difference is notclear. This may be due to lower affinity of rat tissuesthan human tissues to these antibodies that wereraised against human antigens. Further studies areneeded to resolve this important point.

MLE, consisting of multiple layers of squamousand columnar cells, was described by Shields andcolleagues, who reported it in 41% of mucosalbiopsies from the squamocolumnar junction inpatients with BE. All the samples contained gobletcells.26 We found 4/20 cases of MLE at the neo-squamocolumnar junction in the EGDA rats. One ofthe most interesting findings of this study was thediscovery of MLE in the mid-esophagus of twoEGDA rats after careful search in Swiss-roll pre-parations of the esophagus. The low frequency ofMLE in the mid-esophagus may be due to therelatively mild reflux injury to the mucosa com-pared to the distal esophagus. Furthermore, MLE inthe mid-esophagus may go undetected unless serialtissue sections are stained with AB/PAS andexamined carefully. It is likely that MLE has notyet been described in the mid-esophagus of humanfor the same reasons. We speculate that carefulexamination of tissues from BE patients sufferingfrom severe gastroesophageal reflux may revealisolated islands of MLE and scattered mucinouscells high up in the esophagus, away from thegastroesophageal junction.

MLE in humans expresses cytokeratin markers ofboth squamous (CK4 and CK13) and columnarepithelium (CK7 and CK20), suggesting that it maybe a transitional stage between squamous epithe-lium and BE.16,34 It exhibits a certain frequency ofmucin expression (neutral mucin, sulfomucin andsialomucin in 88, 71 and 100%, respectively) andexpresses differentiation markers such as villin andpS2/TFF1. In our study, MLE in the rat model wasfound to share features with human MLE, exceptlesser neutral mucins and no CK20 were expressedin the rat model. Overall, this study suggests thatMLE in the EGDA rat, as in humans, may be anintermediate stage between esophageal squamous

Figure 4 Histochemical and immunohistochemical features of MLE in EGDA rats. (a, b) H&E staining of an area of MLE showingscattered mucinous cells on the top of hyperplastic squamous cells (a). In typical MLE, mucinous cells including columnar and gobletcells are lined up on top of the squamous component (b). (c) AB/PAS staining with hematoxylin counterstaining shows that themucinous cells mainly secrete acid mucins (blue or purple) with little to no neutral mucin (red). (d) HID/AB staining shows that bothsialomucin and sulfomucin are present in the mucinous cells. (e, f) shows CK7 immunoreactivity of the MLE. Note that the mucinouscells in the deep nests of hyperplastic squamous epithelium (e) and surface esophageal mucosa (f) demonstrate strong CK7immunoreactivity. (g) The MLE in rats shows no staining for CK20. (h) shows that a strong cytoplasmic immunostaining of villin is foundin the mucinous cells of MLE in rats. Scale bars: (a–d) 50mm; (e–h) 100mm.


761


Figure 5 Histological characteristics of EAC in EGDA rats. (a) H&E staining shows that EAC consists of dilated, cystic glands withabundant mucin secretion and epithelial dysplasia. The mucinous glands can be seen invading through the muscle layers. (b) showsinvasion of EAC into adjacent liver tissue. Note that residual hepatocytes (r) are seen in the stroma of EAC. (c) shows invasion of EACcells in an adjacent nerve. Note that the invasive cancer cells are breaking through the epineurium and endoneurium of the nerve (shortblack arrow). (d) shows continuity of the EAC (m) with the CLE (-) in the mucosa. (e) HID/AB staining shows large quantities ofsulfomucin (dark brown) and sialomucin (blue) in the extracellular mucinous ‘lakes’ are in the rat EAC. (f) shows that EAC cells stain forCK7 and (g) shows cytoplasmic staining of villin in EAC. Scale bars: (a) 200mm; (b–g) 100mm.


762


epithelium and SIM. The presence of multiple MLEfoci and scattered mucinous cells in the esophagi ofEGDA rats also suggest that SIM may arise fromdifferent clones of the pluripotential stem cells inthe esophagus.

In human patients with BE, SIM and MLE havebeen proposed to arise from progenitor cells inesophageal submucosal glandular ducts or thetransitional cell zone at the gastroesophageal junc-tion.16,35 The rat, however, does not have submuco-sal glands in the esophagus nor does it have thetransitional cell zone between esophageal squamousepithelium and gastric columnar epithelium that areseen in humans. (In the rat the fore-stomach is linedby squamous epithelium.) Instead, EGDA rats have aneo-squamocolumnar junction between esophageal/gastric squamous epithelium and duodenal colum-nar epithelium. It was previously suggested that CLEin the rat model was derived from the creepingsubstitution of esophageal mucosa by the duodenalmucosa.11 However, the histological features andmucin histochemistry in SIM was quite differentfrom the duodenal mucosa. Moreover, the pheno-typic markers of differentiation in the SIM orMLE at the neosquamocolumnar junction and inthe mid-esophagus do not support the view thatthe CLE in the animal model represents creepingsubstitution of the duodenal mucosa. Instead,our studies suggest that SIM in the animalmodel may arise from pluripotential stem cellsin the esophageal squamous epithelium.7 Studiesof gene expression profiles using gene chips mayfurther define the cellular origin of the specializedintestinal metaplasia.

The EGDA rats showed severe changes in thesquamous epithelium consisting of hyperplasia withridge-like surface projections and inverted papillo-ma-like lesions. These lesions did not show featuresof neoplasia. Such lesions have been described inhuman nasal mucosa but not in the human esopha-gus. The reason for this severe squamous epithelialreaction in the EGDA rats but not in human SIM isnot clear. However, it may be related to the moreacute and severe reflux injury in the animal modelthan in clinical conditions in humans.

The EGDA rats also showed changes of dysplasiain the metaplastic epithelium and developed ade-nocarcinomas, as seen in the human disease. Similarto humans, all EACs in the rat were in the distalesophagus. The mucin features, differentiationmarkers such as CK7, Das-1, villin and pS2/TFF1,and tumor markers such as p53 and c-myc in thecells of EAC were similar to those in the adjacentSIM. These findings strongly suggest that EAC inEGDA rats arises from the SIM in the esophagusand is not an unrelated adenocarcinoma arisingfrom the small bowel, as has been suggested by someinvestigators.12

As described earlier, all the EACs in the rat modelwere well-differentiated mucinous adenocarcino-mas.5,10–12 In patients with BE variably differentiatedadenocarcinomas are seen. In our study, only one of11 human EACs was of the mucinous type. Thereason for these differences in the histological typesof adenocarcinomas in the rat and human is notcurrently known. However, this difference may bedue to differences in the expression of certaindifferentiation genes.33

The animal models of esophagoduodenal andesophagogastroduodenal reflux have been used tounderstand the pathogenesis of SIM and EAC. It hasbeen reported that oxidative stress augmented byiron overload is an important risk factors for thedevelopment of EAC in the EGDA rats.36,37 More-over, abnormalities in arachidonic acid have beenreported to be important in the pathogenesis of SIMand EAC.38 A recent study has also demonstratedthe efficacy of COX-2 inhibitors in the preventionof EAC in rats with esophagoduodejejunostomy.39

The present study, showing phenotypic similaritiesbetween the lesions in the EGDA rats and inpatients with BE, suggests that conclusionsdrawn from the animal studies may be relevant toclinical BE.

Similar to humans, SIM in the EGDA rats wasaccompanied by LGD and HGD with a highprevalence of nuclear accumulation of p53 protein(15/20, 75%). Fein et al40 reported that jejunoeso-phagostomy in p53-knockout mice promoted thedevelopment of columnar metaplasia and EAC in

Figure 6 Expression of p53, c-myc and COX-2 in the rat CLE and EAC. (a) shows that CLE in the deep glands exhibit strong nuclear andcytoplasmic staining for p53. (b) shows intermediate cytoplasmic staining of c-myc in areas of dysplasia, and (c) shows a weakcytoplasmic staining of COX-2 in EAC . Scale bars, 50mm.


763


mice.40 It suggests that p53 protein accumulation isan important early event in the pathogenesis of ratEAC, as in humans.

In conclusion, SIM and EAC developed in theEGDA rats are remarkably similar to human BE andEAC. The EGDA rat model may serve as a usefulexperimental system for studying the pathogenesisof BE and the progression of BE to EAC. It has beenshown that concurrent treatment with certainchemical carcinogens promotes development ofsquamous cell carcinoma or adenosquamous carci-noma in the animal model of gastroduodenalreflux.8,9 Further studies in this model may help inunderstanding the molecular pathways that selec-tively lead to one or the other type of esophagealcarcinoma. The animal model may also be helpful intracing changes in signaling pathways involved inthe pathogenesis of esophageal adenocarcinomasand in identifying intermediate markers that maypredict progression of SIM to EAC. In addition, theEGDA rat model can be instrumental for investigat-ing strategies for the prevention and treatment ofhuman Barrett’s esophagus and EAC.37

Acknowledgements

We thank Dr Kiron M Das for providing thepolyclonal antibody Das-1, Dr Howard Chang andDr Lingxin Kong for helpful suggestions.

References

1 Spechler SJ, Goyal RK. The columnar-lined esophagus,intestinal metaplasia, and Norman Barrett. Gastroen-terology 1996;110:614–621.

2 Paull A, Trier JS, Dalton MD, et al. The histologicspectrum of Barrett’s esophagus. N Engl J Med 1976;295:476–480.

3 Shaheen NJ, Crosby MA, Bozymski EM, et al. Isthere publication bias in the reporting of cancer risk inBarrett’s esophagus? Gastroenterology 2000;119:333–338.

4 Falk GW. Barrett’s esophagus. Gastroenterology 2002;122:1569–1591.

5 Fein M, Peters JH, Chandrasoma P, et al. Duodenoeso-phageal reflux induces esophageal adenocarcinomawithout exogenous carcinogen. J Gastrointest Surg1998;2:260–268.

6 Miwa K, Sahara H, Segawa M, et al. Reflux of duodenalor gastro-duodenal contents induces esophageal carci-noma in rats. Int J Cancer 1996;67:269–274.

7 Pera M, Brito MJ, Poulsom R, et al. Duodenal-contentreflux esophagitis induces the development of gland-ular metaplasia and adenosquamous carcinoma in rats.Carcinogenesis 2000;21:1587–1591.

8 Attwood SE, Smyrk TC, DeMeester TR, et al. Duode-noesophageal reflux and the development of esopha-geal adenocarcinoma in rats. Surgery 1992;111:503–510.

9 Pera M, Cardesa A, Bombi JA, et al. Influence ofesophagojejunostomy on the induction of adenocarci-noma of the distal esophagus in Sprague-Dawley ratsby subcutaneous injection of 2,6-dimethylnitrosomor-pholine. Cancer Res 1989;49:6803–6808.

10 Goldstein SR, Yang GY, Curtis SK, et al. Developmentof esophageal metaplasia and adenocarcinoma in a ratsurgical model without the use of a carcinogen.Carcinogenesis 1997;18:2265–2270.

11 Chen X, Yang G, Ding WY, et al. An esophagogastro-duodenal anastomosis model for esophageal adenocar-cinogenesis in rats and enhancement by iron overload.Carcinogenesis 1999;20:1801–1808.

12 Oberg S, Lord RV, Peters JH, et al. Is adenocarcinomafollowing esophagoduodenostomy without carcinogenin the rat reflux-induced? J Surg Res 2000;91:111–117.

13 Boivin GP, Washington K, Yang K, et al. Pathologyof mouse models of intestinal cancer: consensus reportand recommendations. Gastroenterology 2003;124:762–777.

14 Schlemper RJ. International consensus classification ofgastrointestinal epithelial neoplasia: usefulness foresophageal squamous epithelium. In: Imamura M(ed). Superficial Esophageal Neoplasm: Pathology,Diagnosis, and Therapy. Springer-Verlag: Tokyo, Japan,2002, pp 75–82.

15 Rubio CA. Histocytologic criteria for squamous cellcarcinoma in situ of the esophagus. In: Imamura M(ed). Superficial Esophageal Neoplasm: Pathology,Diagnosis, and Therapy. Springer-Verlag: Tokyo, Japan,2002, pp 113–120.

16 Glickman JN, Chen YY, Wang HH, et al. Phenotypiccharacteristics of a distinctive multilayered epitheliumsuggests that it is a precursor in the development ofBarrett’s esophagus. Am J Surg Pathol 2001;25:569–578.

17 Haggitt RC. Barrett’s esophagus, dysplasia, and adeno-carcinoma. Hum Pathol 1994;25:982–993.

18 Jass JR, Filipe MI. The mucin profiles of normal gastricmucosa, intestinal metaplasia and its variants andgastric carcinoma. Histochem J 1981;13:931–939.

19 Stokes ME, Davis CS, Koch GG. Categorical DataAnalysis using the SAS System. SAS Institute, Inc:Cary, NC, 2000.

20 Sheahan DG, Jervis HR. Comparative histochemistry ofgastrointestinal mucosubstances. Am J Anat 1976;146:103–131.

21 Ormsby AH, Vaezi MF, Richter JE, et al. Cytokeratinimmunoreactivity patterns in the diagnosis of short-segment Barrett’s esophagus. Gastroenterology 2000;119:683–690.

22 Das KM, Prasad I, Garla S, et al. Detection of a sharedcolon epithelial epitope on Barrett epithelium by anovel monoclonal antibody. Ann Intern Med 1994;120:753–756.

23 MacLennan AJ, Orringer MB, Beer DG. Identificationof intestinal-type Barrett’s metaplasia by using theintestine-specific protein villin and esophageal brushcytology. Mol Carcinogen 1999;24:137–143.

24 Glickman JN, Wang H, Das KM, et al. Phenotype ofBarrett’s esophagus and intestinal metaplasia of thedistal esophagus and gastroesophageal junction: animmunohistochemical study of cytokeratins 7 and 20,Das-1 and 45 MI. Am J Surg Pathol 2001;25:87–94.

25 Poulsom R, Wright NA. Trefoil peptides: a newlyrecognized family of epithelial mucin-associated mo-lecules. Am J Physiol 1993;265:G205–G213.

26 Shields HM, Rosenberg SJ, Zwas FR, et al. Prospectiveevaluation of multilayered epithelium in Barrett’sesophagus. Am J Gastroenterol 2001;96:3268–3273.

27 Witjes JA, van Balken MR, van de Kaa CA. Theprognostic value of a primary inverted papilloma ofthe urinary tract. J Urol 1997;158:1500–1505.


764


28 Fein M, Fuchs KH, Stopper H, et al. Duodenogastricreflux and foregut carcinogenesis: analysis of duodenaljuice in a rodent model of cancer. Carcinogenesis2000;21:2079–2084.

29 Jenkins GJ, Harries K, Doak SH, et al. The bile aciddeoxycholic acid (DCA) at neutral pH, activates NF-kappaB and induces IL-8 expression in oesophagealcells in vitro. Carcinogenesis 2003, Dec 4 [Epub aheadof print].

30 Tselepis C, Morris CD, Wakelin D, et al. Upregulationof the oncogene c-myc in Barrett’s adenocarcinoma:induction of c-myc by acidified bile acid in vitro. Gut2003;52:174–180.

31 Ireland AP, Peters JH, Smyrk TC, et al. Gastricjuice protects against the development of esopha-geal adenocarcinoma in the rat. Ann Surg 1996;224:358–370.

32 Kawaura Y, Tatsuzawa Y, Wakabayashi T, et al.Immunohistochemical study of p53, c-erbB-2,and PCNA in Barrett’s esophagus with dysplasiaand adenocarcinoma arising from experimental acidor alkaline reflux model. J Gastroenterol 2001;36:595–600.

33 Warson C, Van De Bovenkamp JH, Korteland-Van MaleAM, et al. Barrett’s esophagus is characterized byexpression of gastric-type mucins (MUC5AC, MUC6)and TFF peptides (TFF1 and TFF2), but the riskof carcinoma development may be indicated by the

intestinal-type mucin, MUC2. Hum Pathol 2002;33:660–668.

34 Boch JA, Shields HM, Antonioli DA, et al. Distributionof cytokeratin markers in Barrett’s specialized colum-nar epithelium. Gastroenterology 1997;112:760–765.

35 Jankowski JA, Harrison RF, Perry I, et al. Barrett’smetaplasia. Lancet 2000;356:2079–2085.

36 Chen X, Ding YW, Yang G, et al. Oxidative damage inan esophageal adenocarcinoma model with rats.Carcinogenesis 2000;21:257–263.

37 Chen X, Yang CS. Esophageal adenocarcinoma: areview and perspectives on the mechanism of carci-nogenesis and chemoprevention. Carcinogenesis2001;22:1119–1129.

38 Chen X, Li N, Wang S, et al. Aberrant arachidonicacid metabolism in esophageal adenocarcinogenesis,and the effects of sulindac, nordihydroguaiaretic acid,and alpha-difluoromethylornithine on tumorigenesisin a rat surgical model. Carcinogenesis 2002;23:2095–2102.

39 Buttar NS, Wang KK, Leontovich O, et al. Chemopre-vention of esophageal adenocarcinoma by COX-2inhibitors in an animal model of Barrett’s esophagus.Gastroenterology 2002;122:1101–1112.

40 Fein M, Peters JH, Baril N, et al. Loss of functionof Trp53, but not Apc, leads to the developmentof esophageal adenocarcinoma in mice with jejunoe-sophageal reflux. J Surg Res 1999;83:48–55.


765


phenotype of columnar-lined esophagus in rats with esophagogastroduodenal anastomosis: similarity to...

Documents